Batch Image Downloader — URLs & Pages to KV

Download image URL lists (or extract img / og:image from pages) into Apify KV store with metadata for RAG and archives.

Give this Actor a list of image URLs, HTML page URLs, or both. It fetches each image, stores the bytes in the default key-value store, and writes a dataset row with contentType, bytes, sha256, and dimensions when readable. Failed or empty responses are recorded free by default. Positioning: batch media assets for Agent / dataset pipelines — not spam scraping.

What you get

• Direct image URL batch download into KV (img-<sha256>.<ext> by default)
• Optional page extraction: <img src|srcset>, <picture>, og:image, twitter:image
• Metadata: content type, byte size, SHA-256, width/height (via Pillow when possible)
• Optional WebP / JPEG / PNG re-encode (extra PPE event)
• Duplicate content skip by hash (free duplicate rows)
• Default 512 MB memory; failed / 4xx / empty / non-image free unless you opt in

Measured results

Cloud benches on build 0.1.2, 512 MB (Asia/Taipei 2026-10-02):

Run	Input	Stored	Peak RSS	Platform cost
smoke	3 URLs (2 ok / 1×404 free)	2	87 MB	~$0.00048
bench	12 URLs (10 ok / 2 err free)	10	92 MB	$0.00082 ($0.000082 / image)
pages	w3.org/Standards → 9 imgs	9	93 MB	~$0.00080

Details: docs/PRICING.md.

Use cases

• Build an image asset pack from a sitemap-passed URL list for RAG multimodal pipelines
• Archive product or documentation images with integrity hashes
• Chain after sitemap discovery / URL status checker: only download live image URLs

How to use

{
  "imageUrls": [
    { "url": "https://www.w3.org/Icons/w3c_home.png" },
    { "url": "https://httpbin.org/image/png" }
  ],
  "pageUrls": [],
  "maxImages": 50,
  "maxConcurrency": 6,
  "skipDuplicatesByHash": true
}

Example success row:

{
  "sourceUrl": "https://www.w3.org/Icons/w3c_home.png",
  "status": "success",
  "contentType": "image/png",
  "bytes": 2428,
  "width": 72,
  "height": 48,
  "sha256": "…",
  "kvKey": "img-<sha256>.png",
  "durationMs": 120
}

Pricing (pay-per-event)

Event	Price	When
Actor start	$0.001 / GB (1 GB min)	Once per run
Image downloaded (primary)	$0.0025	Each successful KV store
Image converted	$0.001	Extra when convertFormat succeeds
Failed / empty / non-image / duplicate	$0	Default (chargeFailedImages=false)

Example: 1,000 successful downloads ≈ $2.50 + start (before platform CU/storage). Failed items are never charged by default.

Known limits

• Does not log into sites or bypass paywalls / bot challenges
• SVG may lack width/height; dims come from Pillow for raster formats
• Very large pages: maxImagesPerPage caps extraction
• Copyright / robots: you are responsible for lawful use of downloaded assets
• Not a “scrape any website’s images” spam tool — designed for owned or permitted URL lists

License

AGPL-3.0 (see LICENSE). Runtime stack: httpx (BSD), beautifulsoup4 (MIT), Pillow (MIT-CMU), Apify SDK.